在当前以技术保护为导向的国际背景下，摆脱重要产业领域的技术依赖并全面提升产业技术发展水平是国家的重要战略方向。新兴共性技术是广受关注且尚处于发展初期的领域关键共性技术，也是发展其他创新技术及产品的根基和源泉，对未来产业技术的发展和变革具有较大潜力。识别新兴共性技术是培育共性技术产业链的第一环，在有限的技术资源条件下，可为产业技术研发、攻关和转移转化赢得先机，进而能够率先掌握产业发展的话语权。在现有的新兴共性技术相关研究中，学者们多从定性视角依靠专家智慧进行技术识别，少数从定量视角出发的研究也仅采用单一数据来源，通过专利或论文来对新兴共性技术的特征进行测度。然而单一数据源仅能提供特定的观察视角，很难展现领域技术发展的全貌，这种局限性使得定量分析不够充分和完备，容易导致识别模型产生“数据偏见”。因而，面对当前海量可利用的信息资源，采用多源数据融合的方式，研究适用于不同领域的新兴共性技术识别方法，就成为一个重要课题。

本文在梳理和归纳国内外相关研究成果、探索新兴共性技术内涵特征的基础上，研究基于多源数据融合的新兴共性技术识别方法。本文的主要研究工作如下：1）整合不同渠道的技术信息，选取论文、专利、基金和网络公开信息作为数据来源，构建了包含技术术语、研究人员、组织机构和技术领域的异质复杂网络，从语义和共现两个层面实现了对多源数据的深度融合。2）通过对现有研究工作进行归纳，探索了新兴共性技术的概念和特征，建立了新兴共性技术特征指标体系，为新兴共性技术识别提供了客观的特征测度标准参考。3）基于融合后的多源数据，采用ie-HGCN图卷积神经网络算法，构建了新兴共性技术识别模型。相较于传统基于特征指标或社会网络分析的方法，该模型能够有效地处理高维、非线性的异质复杂网络，更易于捕捉和理解新兴共性技术的抽象特征，从而提升识别结果的准确性。4）本文以“第三代半导体”技术领域为例进行实证研究，结果表明本文提出的方法效果较好，能够细粒度地识别领域新兴共性技术，且识别结果的可解释性较强。

本文的研究工作有助于充分利用海量信息资源，从更客观的角度实现对新兴共性技术的有效识别。同时，本文所提出的方法具有较强的领域适用性，能够推广应用到其他产业领域，基于领域多源数据实现对新兴共性技术的识别，具有一定的学术和应用价值。

In the current international context, which emphasizes technology protection, it is crucial for the country to reduce technological dependence in critical industrial sectors and enhance the overall level of industrial technological development. Emerging generic technologies are key generic technologies that have received widespread attention and are still in the early stages of development. They are also the foundation and source for the development of other innovative technologies, and have great potential for the development and transformation of future industrial technologies. Identifying emerging general-purpose technologies serves as the initial step in nurturing the industrial chain of general-purpose technologies. In the face of limited technological resources, identifying emerging general-purpose technologies provides an advantage in industrial technology research, addressing key challenges and facilitating technology transfer, thereby enabling a leading role in shaping the narrative of industrial development. Existing research in this field has primarily relied on qualitative perspectives and expert opinions for technology identification. However, a limited number of quantitative studies have solely utilized a single data source, such as patents or academic papers, to assess the characteristics of emerging general-purpose technologies. However, relying solely on a single data source limits the observation perspective and makes it challenging to depict the comprehensive landscape of technological development in the field. This limitation renders quantitative analysis inadequate, potentially introducing data biases into the identification models. Therefore, faced with the vast amount of available information resources, it becomes an important research topic to adopt a multi-source data fusion approach, use scientific methods to explore the characteristics and development trends of emerging general-purpose technologies, and study the identification methods applicable to different fields.

 

Based on the review and summary of domestic and foreign research results and the exploration of the connotation and features of emerging general-purpose technologies, this thesis studies the identification method of emerging general-purpose technologies based on multi-source data fusion. The main research work of this thesis is as follows: 1) To incorporate diverse observation perspectives, we choose papers, patents, funding information, and publicly available online data as sources. Subsequently, we construct a heterogeneous complex network that encompasses technical terms, researchers, organizations, and technical fields. Our approach facilitates a comprehensive fusion of multi-source data, considering both semantic and co-occurrence aspects. 2) By summarizing prior research, we elucidate the concepts and defining attributes of emerging general-purpose technologies. Moreover, we develop a feature index system for these technologies, which enables the provision of scientific and objective criteria for their identification. 3) Leveraging the fused multi-source data, we employ the interpretable and efficient Heterogeneous Graph Convolutional Network (ie-HGCN) algorithm to construct an identification model for emerging general-purpose technologies. In comparison to traditional methods relying on feature indicators or social network analysis, this model demonstrates enhanced capacity in handling high-dimensional, nonlinear heterogeneous complex networks. Consequently, it facilitates a more comprehensive comprehension of the abstract characteristics of emerging general-purpose technologies, thereby improving the accuracy of identification results. 4) By conducting empirical research in the "third-generation semiconductor" technology field, we demonstrate the effectiveness of the proposed method in finely identifying emerging general-purpose technologies within the domain. Moreover, the identification results exhibit a high level of interpretability.

 

This thesis significantly contributes to fully harnessing vast information resources and effectively identifying emerging general-purpose technologies from a more objective standpoint. Additionally, the proposed method presented in this thesis exhibits broad applicability across various industry sectors, facilitating large-scale identification of emerging general-purpose technologies within a short timeframe. These contributions substantiate the academic and practical value of the study.

COCCIA M. The source and nature of general purpose technologies for supporting next K-waves: Global leadership and the case study of the US Navy's Mobile User Objective System[J]. Technological Forecasting and Social Change, 2017, 116: 331-339.
郑赛硕, 王学昭, 陈小莉. 共性技术识别方法构建与实证研究——以集成电路行业为例[J]. 图书情报工作, 2021, 65(15): 130-139.
RONDE P. Foresight and Generic Technology: all Exploratory Method with a Knowledge Based Principle to Compare France and Germany[J]. The Netherlands: Eindhoven Centre of Innovation Studies of Eindhoven University of Technology, 2001, 1: 1-42.
虞锡君. 产业集群内关键共性技术的选择——以浙江为例[J]. 科研管理, 2006(01): 80-84.
魏永莲, 唐五湘. 共性技术筛选指标体系及模型研究[J]. 科技管理研究, 2009, 29(04): 46-48.
袁思达. 技术预见德尔菲调查中共性技术课题识别研究[J]. 科学学与科学技术管理, 2009, 30(10): 21-26.
骆正清, 戴瑞. 共性技术的选择方法研究[J]. 科学学研究, 2013, 31(01): 22-29.
张乔木. 基于德尔菲调查和聚类分析的关键共性技术预见研究——以山西省新材料行业为例[J]. 科技管理研究, 2017, 37(13): 121-124.
SMALL H, BOYACK K W, KLAVANS R. Identifying emerging topics in science and technology[J]. Research policy, 2014, 43(8): 1450-1467.
马永红, 孔令凯, 林超然, 等. 基于专利挖掘的关键共性技术识别研究[J]. 情报学报, 2020, 39(10): 1093-1103.
樊霞, 陈双丽. 产学研合作更有利于产业共性技术研发创新吗——对USPTO中国专利数据的实证检验[J]. 科技进步与对策, 2016, 33(13): 59-65.
郑彦宁, 浦墨, 刘志辉. 基于产业创新链的产业共性技术识别基本理论探讨[J]. 情报理论与实践, 2016, 39(09): 53-58.
覃兴. 基于专利分析的产业共性技术识别探讨[J]. 产业与科技论坛, 2017, 16(05): 62-63.
王颖, 孙成权. 技术预见的典型模式分析[J]. 图书与情报, 2008(2): 63-66.
NAGY D, SCHUESSLER J, DUBINSKY A. Defining and identifying disruptive innovations[J]. Industrial marketing management, 2016, 57: 119-126.
王昆声, 周晓纪, 龚旭, 等. 中国工程科技2035技术预见研究[J]. 中国工程科学, 2017, 19(1): 34-42.
NYIRI L. Foresight as a Policy-making Tool[J]. Technology Foresight for Organizers, Ankara, Turkey, December, 2003, 1: 1-12.
JÄRVENPÄÄ H M, MÄKINEN S J, SEPPÄNEN M. Patent and publishing activity sequence over a technology's life cycle[J]. Technological Forecasting and Social Change, 2011, 78(2): 283-293.
PARK I, YOON B. Technological opportunity discovery for technological convergence based on the prediction of technology knowledge flow in a citation network[J]. Journal of Informetrics, 2018, 12(4): 1199-1222.
曹兴, 孙绮悦. 新兴技术知识网络跨界融合的机理与实证研究[J]. 科学学研究, 2021, 39(05): 854-866.
化柏林, 李广建. 大数据环境下多源信息融合的理论与应用探讨[J]. 图书情报工作, 2015, 59(16): 5-10.
张维冲, 王芳, 赵洪. 多源信息融合用于新兴技术发展趋势识别——以区块链为例[J]. 情报学报, 2019, 38(11): 1166-1176.
张小艳, 白瑜. 基于加权融合字词向量的中文在线评论情感分析[J]. 计算机应用研究, 2022, 39(01): 31-36.
COSTAS R, DE RIJCKE S, MARRES N. “Heterogeneous couplings”: Operationalizing network perspectives to study science‐society interactions through social media metrics[J]. Journal of the Association for Information Science and Technology, 2021, 72(5): 595-610.
LYU X, COSTAS R. How do academic topics shift across altmetric sources? A case study of the research area of Big Data[J]. Scientometrics, 2020, 123(2): 909-943.
徐路路, 王效岳, 白如江. 基于PLDA模型与多数据源融合相关性分析的新兴主题探测研究——以石墨烯领域为例[J]. 情报理论与实践, 2018, 41(4): 63-69, 43.
SUGIMOTO C R, WORK S, LARIVIÈRE V, et al. Scholarly use of social media and altmetrics: A review of the literature[J]. Journal of the association for information science and technology, 2017, 68(9): 2037-2062.
ORTEGA J L. How do media mention research papers? Structural analysis of blogs and news networks using citation coupling[J]. Journal of Informetrics, 2021, 15(3): 101175.
卢小宾, 翟兴, 王涛. 基于复杂网络的信息分析模型研究[J]. 情报理论与实践, 2020, 43(05): 177-182.
江娴, 魏凤. 基于专利分析的共性技术识别研究框架[J]. 情报杂志, 2015, 34(12): 79-84.
张鹏, 杨艳君, 宋丽雪. GPS产业的共性技术识别、演进及其启示——基于专利分析[J]. 技术经济, 2016, 35(12): 60-75, 89.
贺正楚, 张蜜, 吴艳, 等. 生物医药产业共性技术服务效率研究[J]. 中国软科学, 2014(2): 130-139.
朱桂龙, 朱明晶, 尹潇. 知识扩散视角下共性技术的商业化评价——基于多层网络的反向识别方法[J]. 科学学与科学技术管理, 2019, 40(2): 16-25.
吴菲菲, 李倩, 黄鲁成. 基于专利 SAO 结构的技术应用领域识别方法研究[J]. 科研管理, 2014, 35(6): 1-7.
TANG L, PENG S, BI Y, et al. A new method combining LDA and PLS for dimension reduction[J]. PloS one, 2014, 9(5): e96944.
ZHANG Y, WU M, MIAO W, et al. Bi-layer network analytics: A methodology for characterizing emerging general-purpose technologies[J]. Journal of Informetrics, 2021, 15(4): 101202.
田鹏伟, 张娴. 基于异构信息网络融合的专利技术主题识别研究[J]. 情报杂志, 2021, 40(08): 45-52.
LI B, PI D, LIN Y, et al. Multi-source information fusion based heterogeneous network embedding[J]. Information Sciences, 2020, 534: 53-71.
WANG R W, WEI S X, FRED Y Y. Extracting a core structure from heterogeneous information network using h-subnet and meta-path strength[J]. Journal of Informetrics, 2021, 15(3): 101173.
黄立威, 李德毅, 马于涛, 等. 一种基于元路径的异质信息网络链路预测模型[J]. 计算机学报, 2014, 37(04): 848-858.
PENG T Q. Assortative mixing, preferential attachment, and triadic closure: A longitudinal study of tie-generative mechanisms in journal citation networks[J]. Journal of Informetrics, 2015, 9(2): 250-262.
吴颖文, 纪杨建, 顾新建. 基于专利动态复杂网络的产业共性技术预测[J]. 计算机集成制造系统, 2020, 26(12): 3185-3194.
DAVID P A. Computer and dynamo: The modern productivity paradox in a not-too distant mirror[R]. 1989, 1: 315-348.
BRESNAHAN T F, TRAJTENBERG M. General purpose technologies ‘Engines of growth’?[J]. Journal of econometrics, 1995, 65(1): 83-108.
JOVANOVIC B, ROUSSEAU P L. General purpose technologies[M]//Handbook of economic growth. Elsevier, 2005, 1: 1181-1224.
周密. 我国自主创新的结构性问题与困境摆脱[J]. 改革, 2009(05): 98-102.
李纪珍. 产业共性技术供给体系[M]. 北京: 中国金融出版社, 2004.
许端阳, 徐峰. 产业共性技术的界定及选择方法研究——基于科技计划管理的视角[J]. 中国软科学, 2010(04): 73-79.
栾春娟. 战略性新兴产业共性技术测度指标研究[J]. 科学学与科学技术管理, 2012, 33(02): 11-16.
张健, 张威, 吴均. 战略性新兴产业共性技术协同创新的演化博弈——三重螺旋视阈下的研究[J]. 企业经济, 2017, 36(01): 41-48.
AITHAL P S, AITHAL S. Management of Universal Technologies & their Industry Implications[C]//Proceedings of International Conference on Emerging Trends in Management, IT and Education. 2019, 1(2): 318-328.
GOLDFARB A, TASKA B, TEODORIDIS F. Could machine learning be a general-purpose technology? Evidence from online job postings[J]. SSRN Electronic Journal, 2021, 1(2): 1-66.
于佳会, 刘佳静, 郑建明. 多源多维数据融合研究态势：理论、方法与应用[J]. 情报杂志, 2022, 41(05): 133-138+207.
DEVLIN J, CHANG M W, LEE K, et al. Bert: Pre-training of deep bidirectional transformers for language understanding[J]. arXiv preprint arXiv:1810.04805, 2018, 1(1): 4171-4186.
石川, 王睿嘉, 王啸. 异质信息网络分析与应用综述[J]. 软件学报, 2022, 33(02): 598-621.
邱均平. 文献计量学[M]. 北京: 科学技术文献出版社, 1988.
KIPF T N , Welling M . Semi-supervised classification with graph convolutional networks[J]. arXiv preprint arXiv:1609.02907, 2016.
张娴, 方曙, 肖国华, 等. 专利文献价值评价模型构建及实证分析[J]. 科技进步与对策, 2011, 28(6): 127-132.
王敬东, 张智雄. 基于逆序扫描和共现分析的缩略语快速提取算法[J]. 计算机应用研究, 2018, 35(03): 700-704.
罗威, 武帅, 田昌海. 数据驱动的技术预测之研究评析——以FUSE项目为例[J]. 情报理论与实践, 2019, 42(07): 15-19+34.
ZHANG Y, AI Q, CHEN X, et al. Joint representation learning for top-n recommendation with heterogeneous information sources. In: Proc. of the 2017 ACM on Conf. on Information and Knowledge Management. 2017: 1449-1458.
LI E Y, LIAO C H, YEN H R. Co-authorship networks and research impact: A social capital perspective[J]. Research Policy, 2013, 42(9): 1515-1530.
张金柱, 李溢峰. 专利分类序列和文本语义表示视角下的技术融合预测研究[J]. 情报学报, 2022, 41(06): 609-624.
李纲, 余辉, 毛进. 基于多层语义相似的技术供需文本匹配模型研究[J]. 数据分析与知识发现, 2021, 5(12): 25-36.
刘小平, 冷伏海, 李泽霞. 国际科技前沿分析的方法和途径[J]. 图书情报工作, 2012, 56(12): 60-65.
CHRISTENSEN C M. The innovator's dilemma: when new technologies cause great firms to fail[M]. Harvard Business Review Press, 2013.
CARLEY S F, NEWMAN N C, PORTER A L, et al. An indicator of technical emergence[J]. Scientometrics, 2018, 115(1): 35-49.
KOPPENS F H L, MUELLER T, Avouris P, et al. Photodetectors based on graphene, other two-dimensional materials and hybrid systems[J]. Nature nanotechnology, 2014, 9(10): 780-793.
Yole Intelligence. GaN RF Market: Applications, Players, Technology, and Substrates 2021[R/OL]. https://www.yolegroup.com/product/report/gan-rf-market-applications-players-technology-and-substrates-2021.html, June, 2021
李一村, 郝晓斌, 代兵, 等. MPCVD单晶金刚石高速率和高品质生长研究进展[J]. 人工晶体学报, 2020, 49(06): 979-989.
王艳丰, 王宏兴. MPCVD单晶金刚石生长及其电子器件研究进展[J]. 人工晶体学报, 2020, 49(11): 2139-2152.
王志越, 巩小亮, 付丙磊. 扬帆芯未来，助力新基建，大力推进第三代半导体制造装备国产化[J]. 科技导报, 2021, 39(14): 83-91.
LIU Y, FANG Y, YANG D, et al. Recent progress of heterostructures based on two dimensional materials and wide bandgap semiconductors[J]. Journal of Physics: Condensed Matter, 2022, 34(18): 1-20.
刘国友, 王彦刚, 李想, 等. 大功率半导体技术现状及其进展[J]. 机车电传动, 2021(05): 1-11.
BADER S J, LEE H, CHAUDHURI R, et al. Prospects for wide bandgap and ultrawide bandgap CMOS devices[J]. IEEE Transactions on Electron Devices, 2020, 67(10): 4010-4020.
彭洋, 陈明祥, 罗小兵. 深紫外LED封装技术现状与展望[J]. 发光学报, 2021, 42(04): 542-559.